Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Alpaca: A strong open-source instruction-following model

crfm.stanford.edu

Post reply on HN
Print view View original

201–210 of 313 posts

Re: Alpaca: A strong open-source instruction-following model

#201

Post by jxy »

LeoPanthera wrote: post #194
pmoriarty wrote: post #172

Earlier quoted context omitted.

If it's indistinguishable, then how does it matter who/what it was written by?

Because I am far less likely to trust encyclopedic content written by an AI, which can lie so convincingly that it does not know that it is lying. I have even seen it generating fictional source references that are hard to verify. I just don't trust it.

Maybe we just need a better attribution, so people can decide on the level of objective confidence. If you live in the US, it is already difficult to avoid seeing and listening to real people lie publicly, which may take journalists a long time to verify. Put it in another way, we might just see late night talk show hosts start parodying $next_gen_llm instead of real people, if this $next_gen_llm started affecting people in a more profound way.

Re: Alpaca: A strong open-source instruction-following model

#202

Post by O__________O »

haolez wrote: post #197

Do we have some studies on the impact of extreme quantization on these models? In my understanding, the efforts to make them run in commodity hardware involves the quantization of the weights to 8 bits or even 4 bits. Is the model still reliable after this?

Yes, intro on the topic if anyone is interested, include topic of benchmarking degradation:

- https://huggingface.co/blog/hf-bitsandbytes-integration

Re: Alpaca: A strong open-source instruction-following model

#203

Post by sebzim4500 »

stavros wrote: post #183
sebzim4500 wrote: post #139

Earlier quoted context omitted.

I haven't seen anything official from OpenAI confirming that ChatGPT has fewer than 175B parameters, although it is a reasonable guess if you read between the lines of their statements. Given the author of that article is a CEO of an 'AI Ad Optimization Platform' I think that number is speculative at best.

ChatGPT is fine-tuned InstructGPT, which has 1.3B parameters, if I'm not mistaken. Reference for the former: https://www.technologyreview.com/2023/03/03/1069311/inside-s...

InstructGPT isn't a single model, it's a set of techniques for finetuning a foundation model

Re: Alpaca: A strong open-source instruction-following model

#204

Post by dang »

nlowell wrote: post #119
dang wrote: post #110

Earlier quoted context omitted.

Please don't take HN threads further into ideological flamewar. It's not what this site is for, and we have to ban accounts that do it repeatedly. If you'd please review https://news.ycombinator.com/newsguidelines.html and stick to the rules when posting here, we'd appreciate it.

Sorry if I made it worse, I really felt like people's opinions were being treated uncharitably, and I was trying to right it.

I definitely appreciate the wish to treat people charitably!

In terms of the site guidelines, "You're missing the point" is kind of a swipe and so should probably be dropped; "willfully" should definitely have been dropped because it's making a claim about negative intent that one can't actually know and such claims always land as an attack on the other person; and the last sentence was snarky and should have been dropped.

If one makes a habit of editing such things out of one's comments, one's substantive point will come to the fore more clearly, which benefits everyone. But it's not always easy in the moment!

Re: Alpaca: A strong open-source instruction-following model

#205

Post by dang »

beepbooptheory wrote: post #117
dang wrote: post #111

Earlier quoted context omitted.

Please don't take HN threads further into ideological flamewar. It's not what this site is for, and we have to ban accounts that do it repeatedly. If you'd please review https://news.ycombinator.com/newsguidelines.html and stick to the rules when posting here, we'd appreciate it.

Ah! Yes, sorry, this is fair Dang. Happy to die on this hill today, as it were, will do better in future though. cheers.

Appreciated!

Re: Alpaca: A strong open-source instruction-following model

#206

Post by sanxiyn »

isaacfung wrote: post #57
bilsbie wrote: post #48

What do they mean by instruction? Is it just regular LLM?

LLM just predicts the next token given the previous tokens(this can be trained without manual labelling by humans). Instruct GPT and ChatGPT use reinforcement learning from human feedback to align the model with human intents so it understands instructions. https://huggingface.co/blog/rlhf

Note that Alpaca is NOT using RLHF. It explicitly states it used supervised finetuning.

Re: Alpaca: A strong open-source instruction-following model

#208

Post by vishal0123 »

scottlawson wrote: post #152
redox99 wrote: post #29

They used OpenAI davinci to generate the dataset for instruction tuning. They mention one of the reasons this model is noncommercial is because OpenAI's terms of use prohibit developing models that compete with OpenAI. They also published the generated dataset. If one were to use this dataset to tune your own commercial model, would this be OK, or illegal? I would guess OpenAIs ToS no longer applies, as this is now A…

This model is for non commercial use only so it's unclear to me how it could compete commercially with OpenAI. Sure, you could decide to interpret "compete" in a way that makes it seem like this violates TOS, but the mom commercial restriction is a powerful counterargument too.

They allow others to use it for noncommercial use. Other research groups won't have to use openAI APIs for some of the usecases, hence the model is competing.

Re: Alpaca: A strong open-source instruction-following model

#209

Post by tempaccount420 »

technerder wrote: post #148

Is there anyway to use Google's Coral TPU PCIe Accelerator[1]'s to increase the inference performance of GPT models like this and Llama? [1] https://coral.ai/products/pcie-accelerator/

Supposedly TPUs are better than GPUs, but everyone other than Google uses GPUs for models.

Re: Alpaca: A strong open-source instruction-following model

#210

Post by make3 »

vishal0123 wrote: post #207

I am surprised that they were able to launch this in stanford's domain. They clearly broke TOS of both Facebook and OpenAI, and even admitted doing that. I would be happy if the research decides to ignore openAI and facebook's useless restrictions.

doing that would make it less likely that they would publish a model next time
Post reply on HN